About acrowther

acrowther · ‎07-19-2017

That did it - thank you!!!

acrowther · ‎07-19-2017

Hi, I have a dataset that contains three ID variables: FieldA, Field, FieldC. An individual's ID can appear in one, two, or all three of these fields. Also, the number of observations per variable varies (i.e. there are different numbers of obs in FieldA, FieldB,and Field C). Below is a snapshot of the data (there are thousands of records overal). FieldA FieldB FieldC 100128843 100028331 100028331 100153363 100038471 100038471 100265327 100042435 100042435 100281941 100046405 100046405 100533731 100055129 100055129 100590741 100062169 100062169 100682161 100065289 100065289 100691903 100071483 100071483 100696993 100071671 100071671 100780833 100078509 100078509 100874907 100083405 100083405 100875239 100084351 100084351 I'm trying to determine how many IDs appear in: Only FieldA Only FieldB Only FieldC Only FieldA + FieldB Only FieldA + FieldC Only FieldB + FieldC Only Field A + FieldB + Field C I'd like for these groups to be mutually exclusive. Any advice? Thanks!

acrowther · ‎02-03-2017

Thanks for replying. I ran this code - seems to give a distribution of the most frequent DX2 codes across all observations (and so on for DX3-30). I think the next step is figuring out which DX values are the mostcommon across DX2-30 (not just within each DX). Does that make sense?

acrowther · ‎02-03-2017

Great question. Yes, the Top 10 DX refers to counts from all DXs combined (DX2-DX30). They can definitely be clubbed for counting purposes.

acrowther · ‎02-03-2017

Thanks for your replies. Below is the datastep (if that helps; I can provide more detailed data if needed) I used to create my data and also the chart I'd like to be able to populate. DX1 is used as the initial filter to create the derrivative dataset. From that secondary dataset (workdata.discharge_5990), I'd like to be able to determine the most frequent DX codes that appear in DX2 - DX30 for all observations, rather than just a single row. The dataset "workdata.discharge_5990" has more than 100,000 observations, if that helps guide the approach. It does not have a variable that functions as a unique identifier though all the observations are unique. Top Ten Most Common DX for Paients with DX1 of 5990 N % DX A DX B DX C DX D DX E DX F DX G DX H DX I DX J data workdata.discharge_5990; set workdata.emergency_department; where dx1 = '5990'; run;

acrowther · ‎02-03-2017

Hi all: The question I need to answer is: for people who have a defined DX1 value (in this case, 5990), what are the most frequent codes that exist across DX2 through DX30? I'm trying to build out a profile of the "most common case" based on these codes. In other words, for people with 5990, the ten most frequent codes are "A (freq, x%) ,B (freq, x%) etc..." I think I need two steps: Create a dataset containing observations with 5990 in DX1. This is done already. For DX2 - DX 30 (the dataset extends through 30 but I trucated to make it easier to see), determine the most frequent values across all 29 variables and across all other observations. Many thanks for your help! (I use 9.4, if that helps.) DX1 DX2 DX3 DX4 DX5 DX6 DX7 DX8 5990 2859 1302 V5869 5990 5990 2768 42789 49390 25000 2720 V5869 V5866 5990 4149 25000 4019 V5866 V5869 5990 5990 4019 5715 45621 5990 7881 4019 2720 53081 V5869 V4589 V5866 5990 42731 25000 V5869 5990 27650 V5869 53081 4019 2720 5990 78701 78791 5990 4019 2720 V5869 5990 78820 78097 29420 V1309 5990 413 4264 78607 V1254 33829 5990 40391 5856 V4511 V4573 V1051 1970 V1011

Online Status	Offline
Date Last Visited	‎07-20-2017 10:43 AM

Re: How do I define and count mutually exclusive groups?

How do I define and count mutually exclusive groups?

Re: How do I count observations across variables?

Re: How do I count observations across variables?

Re: How do I count observations across variables?

How do I count observations across variables?

Re: How do I define and count mutually exclusive groups?

How do I define and count mutually exclusive groups?

Re: How do I count observations across variables?

Re: How do I count observations across variables?

Re: How do I count observations across variables?

How do I count observations across variables?